Improving the Effectiveness of Stress Tests: Strategies for Greater Harmonization
Jo Paisley, Co-PresidentGARP Paris Chapter, 6 June 2019
Bank shares plunge across Europe as stress tests warn of contagion
the guardian 3 August 2016RBS biggest failure in Bank of England stress test30 November 2016NEW
S
B B C
USA TODAY
Citigroup sinks 5% after failing Fed 'stress test’
Stress Testing Headlines
2007 Context
• Worst recession since Great Depression• Freezing up of many parts of the financial markets• Capital ratios were not forward-looking• Markets unable to distinguish weak from healthy banks
• Basel (2009) showed severe weaknesses in banks’ stress testing• Firms that were stress testing well performed better.• Basel published useful set of principles to guide banks and supervisors• Regulators started to use stress tests in earnest
3
Regulators’ Regimes Differ
Horizon and frequency
9 quartersEach year (plus DFAST)35 banks
Annual cyclical scenario 5 yearsbiennial exploratory 7 years7 banks
Biennial, 3 years scenario horizon
48 banks
Capital hurdle rate Capital requirements based on 4.5% CET1 threshold New stress test capital buffer Approval /objection of planned dividends or share buybacks
System wide macro-prudential plus input into microprudential buffer settingHurdle rate reflects individual P2A, complicated by IFRS9New capital plan may be required
No threshold used Minimum of 5.5% soft hurdle used for P2G setting P2G does not trigger MDA restrictions
Focus Global macroeconomic scenario with US focus
ACS is a global macroeonomicdownturn; traded risk stress plus misconduct stress. BES set to cover ‘dark corners’
Global with European focus
Key methodology assumptions
Dynamic balance sheet Dynamic balance sheet with mandated UK lending paths
Static balance sheetMultiple EBA mandated constraints (eg on NII)
Global Banks Face a Myriad of Tests
5
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
CCAR (HNAH) DFAST (HNAH)
EBA / ECB Stress Testing - ALL
Argentina ICAAP (starts Dec)
Brazil ICAAP
Uruguay ICAAP
Mexico Local Regulatory ST exercises
Mexico Brokerage House Stress Testing (new in 2016)
HBME Regional ST (Interest Rate Shock) HBME & UAE Regional stress test ( EM shock and low oil price refresh)
HBME Large Exposure and Reverse stress test
UAE reverse stress test
HBON: Liquidity stress test, 3 capital stress tests (solvency, resilience and concentration risk) reporter quarterly to CBO
Qatar: Half yearly regulatory mandated
Group Resolution & Recovery Plan (TBC) - ALL
Internal: ad hoc internal stress tests for specific regional portfolios requested by senior management
EBA/ECB ST – HBFR & HBMT Standalone submission Recovery & Resolution Plan –HBEU Solo, HBFR, UK RFB
PRA Stress Testing - ALL Group ICAAP - ALLInternal ST (TBC) - ALL
1. Quarterly Regional Specific scenarios (HBEU Solo & Cons) / 2. Additional county level analysis for HBFR, HBMT, HBTR & Board
HKMA SDST Singapore Recovery Plan
China
India
Taiwan / ThailandSri Lanka
Australia Reverse STSingapore MAS IWST
Philippines
MENA: Periodic stress testing as required at site level (Ad-hoc)
Australia APRA HKMA RST
Malaysia BNM supervisory Stress Test
Singapore Reverse Stress Testing
PRA ST – HBEU Standalone Submission Internal Canadian stress testing activities
Group
US
Latin America
Middle East
Europe
Asia
Canada
Source: Trevor Wells presentation to RiskMinds December 2017
16.214.6 14 13.3
14.5
9.9 9.4 8.67.3
8.89.6
5.8 6.3 6.9 7.29.7 9.1
11.4 11 10.3
RBS HSBC LLOYDS BARCLAYS AVERAGE
CET1 capital ratios: pre and post stress
2017 CET1 ratioEBA post-stress CET1 ratioBoE post-stress CET1 ratio (before management actions)BoE post-stress CET1 ratio (after management actions)
Comparing stress tests: EBA and BOE 2018
Benefits vs. Costs
Potential benefits are substantial
‣ Forward-looking
‣ Insights on firms/the system
‣ Crisis management tool
‣ Market discipline
‣ Data quality improvements
‣ Improves public confidence
But so are the costs…
7
Costs of Multiple, Disparate Stress Tests
‣ Regulatory tests hard to compare
‣ Publication standards differ across regulators – confuses investors
‣ Crowds out banks’ internal stress testing
‣ Encourages a ‘compliance’ mindset
‣ Firms discouraged from investing in robust, strategic IT architecture
‣ Can’t achieve economies of scale in production of stress test outputs
‣ Capital levels Banks business models may not be
taken into account Diversification effects at Group not
taken into account Risk of double counting capital for
micro/macro prudential purposes
8
‣ Published in December 2018
‣ Kickstart dialogue between practitioners and regulators
Basel Principles Provide the Overarching Framework
‣ Have clear objectives, effective governance
‣ Be used as a risk management tool to inform business decisions
‣ Cover material and relevant risks, be sufficiently severe
‣ Have sufficient resources
‣ Be supported by sufficiently granular data and robust IT systems
‣ Be based on fit for purpose models/methodologies
‣ Be reviewed and challenged
‣ Practices/findings should be communicated within and across jurisdictions
10
Anatomy of a Supervisory Stress Test
11
INPUT TEMPLATES
SCENARIOS OUTPUT TEMPLATES
DISCLOSURE
Reporting templates
Portfolio structure & granularity
Definitions
Accounting standards
Key assumptions and methodology
Risk coverage
Scenario and variables
Time horizon
Public/private disclosure
Use of results
Management actions
Disclosure templates
PURPOSE PROCESS COMPONENTS1 32
Home Host protocol
Scenario selection
Timetable
GLOBAL GUIDELINES
Proportionality & consultation
Micro prudential e.g. Capital/Liquidity planning
Macro prudential e.g. System-wide resilience
Scenarios can by common or set by
individual regulators.Aim to harmonise how
they are specified.
The Purpose
‣ The Purpose will drive elements of the process‣ Design should be compatible its stated purpose.‣ Basel Principles:
1 Objectives2 Governance8 Challenge and review
PURPOSE1
Micro prudential: e.g. Capital/
Liquidity planning
Macro prudential:
e.g. System-wide
resilience
Global Guidelines
‣ Should apply to all supervisory stress tests.‣ Promote harmonisation. ‣ Basel Principles:
3 Use test4 Scenario suitability5 Adequacy of resources6 Data and IT
2
Home Host protocol
Scenario selection
Timetable
GLOBAL GUIDELINES
Proportionality & consultation
Global Guidelines
Only home supervisors can run group-wide stress tests.
Home Host protocol
Scenario selection
Timetable
Proportionality & consultation
2 GLOBAL GUIDELINES
Global Guidelines
Only home supervisors can run group-wide stress tests.
Can jurisdictions coordinate timing? Given firms sufficient notice. Balance supervisory-run vs internal stress tests.
Home Host protocol
Scenario selection
Timetable
Proportionality & consultation
2 GLOBAL GUIDELINES
Global Guidelines
Only home supervisors can run group-wide stress tests.
Can jurisdictions coordinate timing? Given firms sufficient notice. Balance supervisory-run vs internal stress tests.
Think costs/benefits and ‘optimal’ data granularity. Reduce intensity for immaterial portfolios/risks
Home Host protocol
Scenario selection
Timetable
Proportionality& consultation
2 GLOBAL GUIDELINES
Global Guidelines
Only home supervisors can run group-wide stress tests.
Can jurisdictions coordinate timing? Given firms sufficient notice. Balance supervisory-run vs internal stress tests.
Think costs/benefits and ‘optimal’ data granularity. Reduce intensity for immaterial portfolios/risks
Scenarios must be suitable for the stated goalsConsider severity of scenarios across jurisdictions.
Home Host protocol
Scenarioselection
Timetable
Proportionality & consultation
2 GLOBAL GUIDELINES
Process Components
INPUT TEMPLATES
SCENARIOS OUTPUT TEMPLATES
DISCLOSURE
Reporting templates
Portfolio structure & granularity
Definitions
Accounting standards
Key assumptions and methodology
Risk coverage
Scenario and variables
Time horizon
Public/private disclosure
Use of results
Management actions
Disclosure templates
PROCESS COMPONENTS3
Scenarios can by common or set by
individual regulators.Aim to harmonise how
they are specified.
Methodology examples:‣ Static/dynamic balance sheet‣ Treatment of discontinued
businesses ‣ Foreign currency conversion‣ Holding periods for traded
risk shocks.
Summary‣ Good time to reassess the approach to supervisory stress tests. ‣ Code of Practice to kick start dialogue‣ Be clear on purpose‣ Global guidelines provide common ‘ground rules’. ‣ Process standards should encourage:
• Co-ordination across regulators. • Standards on stress test execution.• Standards on consultation. • Harmonisation, where appropriate. • A materiality and cost/benefit approach.